Google 发布 Gemini 3.8 Live 两款音频模型

Google DeepMind 上周推出 Gemini 3.8 Live 与 Live Extended Thinking 两款实时音频模型，官方展示创意、客服与机器人场景用法。

Google

@Google2026.09.22收录原推发布于 2026.09.22

Google @Google原推发布于 2026.09.22

在 X 上打开原推

Last week, we said “hello” to two new audio models from @GoogleDeepMind: Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. See how members of our team are building with them across creative applications, real-time support, and robotics:

x.com/google/status/2102146240968270047采集于 2026.09.22 10:36

在 X 上打开原推